metrique 0.3.0-2a

Metrique - Client Libraries

Metrique

Python/MongoDB Data Warehouse and Information Platform

metrique provides a simple python API to support ETL workloads for extracting data from disperate sources, iteratively, rapidly and reproducibly, with transparent, historical object persistence and tight clientside integration with popular python scientific computing libraries to faciliate creation and publication of a wide variety of analysis and reports, large and small.

Author: “Chris Ward” <[e-mail]> Sources: https://github.com/kejbaly2/metrique

Quick Install (auto-deploy -> virtenv)

The instructions given below assume fedora rpm package names:

# prerequisite *os* packages
sudo yum install python python-devel python-setuptools python-pip
sudo yum install git gcc gcc-c++ gcc-gfortran
sudo yum install freetype-devel libpng-devel # matplotlib deps

sudo yum install mongodb mongodb-server

# additional python global dependencies, from pip
sudo pip install pip-accel  # faster cached pip installed

# make sure our core package managers are up2date
sudo pip-accel install -U distribute setuptools

# our installation directory should always be a py virtualenv
sudo pip-accel install virtualenv

# get metrique sources
git clone https://github.com/kejbaly2/metrique.git
cd metrique

# deploy metrique master branch into a virtual environment
# including dependencies.
# NOTE this takes 10-15 minutes to compile everything from source!
./metrique.py -V ~/metrique.master deploy --all

# activate the virtual environment
source ~/metrique.master/bin/activate

# to persist data, start mongodb
./metrique mongodb firstboot
./metrique mongodb start

# launch ipython, connect to a metriqued instance and start mining!